Having trouble replicating the exact Rheumatoid Arthritis example

[Salik Syed]

Hi All,

We have downloaded the spreadsheet linked in the nature paper (nature13835-s1.xls) -- see here:

https://www.nature.com/nature/journal/v518/n7539/extref/nature13835-s1.xls

I looked at the SNPs for Rheumatoid arthritis and found that the spreadsheet contains 195 of them (not 84 as shown in the demo video + paper).

I use the same steps as outlined here: https://github.com/ryanlayer/giggle#example-analysis

(except my input BED file just contains the location of SNPs as 1 base pair long intervals)

When I run the analysis I get a very different heatmap than the one shown in the paper. What am I missing? Any help is much appreciated.

Best,

Salik

[Ryan Layer]

Try this one 

https://github.com/ryanlayer/giggle/tree/master/experiments/gwas

--
You received this message because you are subscribed to the Google Groups "giggle" group.
To unsubscribe from this group and stop receiving emails from it, send an email to giggle-discus...@googlegroups.com.
To post to this group, send email to giggle-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/giggle-discuss/CAPxOnbJQv_s7kdPxRx3eVnjhYitmAyUsGEns5cNK9X_5aFEjRA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Salik Syed]

Thanks for the response Ryan! Turns out we were building the index wrong (we ran liftover on the BED data to match the SNPs but did not realize it was already in GRCh38) after correcting this the analysis worked seamlessly!

To view this discussion on the web visit https://groups.google.com/d/msgid/giggle-discuss/CALN0tX%2BuyxjU_E_aatRPqfyQcVNevPxAnsvD%2BzVYgJEaO%2BfwRA%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

Sent from my iPhone

[Ryan Layer]

Great! Ya for reproducibility! 

[Ryan Layer]

On Sun, Oct 28, 2018 at 8:56 PM Yudong Qiu <q...@me.com> wrote:

Hi Ryan,

Thanks a lot for the reply last time. After being able to reproduce the example, we tried to build a new giggle index (based on the original one) and to use it in searching with queries. However, we found that the `giggle search` command got stuck and allocate more and more memory until got killed by the operating system. This happens when searching against the newly built index but not against the original index from the example. We are running on Ubuntu 16.04 with giggle built by calling `make` in the “giggle" folder.

Here are some steps and details of what we did:

1. We filtered the intervals in each bed.gz file in the `roadmap_sort` folder which was used to create the original giggle index in the example. We kept the same name of the bed files, and found the resulting bed.gz files are significantly smaller than the original ones. Some of them become empty after filtering and we tried to keep or remove the empty ones.

2. We used the command `giggle index -I “bed_files/*.gz” -o index` to build the new giggle index and it finished successfully.

Assuming your files are sorted, try adding "-s -f"

 

3. We used the command `giggle search -I index -q processed.bed.gz -s` to search with the query file `processed.bed.gz` against the newly built index in `index/`, the issue occurs that the giggle program got stuck with no output, and memory usage keep increasing quickly until used up all 64G then killed by OS.

4. We used the command `giggle search -I roadmap_sorb_b -q processed.bed.gz -s` to search with the same query `processed.bed.gz` against the original index `roadmap_sort_b`, and it works successfully.

To be able to reproduce this issue, I compressed the files related in this issue. In the attached `custom_index.tar.gz`, after decompressing, in the `./custom_index/with_empty` folder, are the `bed_files` folder which contains all bed files used in creating the index; The `index` folder containing the resulting index; And the `processes.bed.gz` file is the query bed file. I also tried to remove the empty bed files before creating the giggle index, and the tests are done in the `./custom_index/empty_bed_removed` folder with the same file names.

Any hint on why the `giggle search` command  would freeze and keep allocating memory infinitely would be very helpful for us. We greatly appreciate your help!

Best,

Yudong & Salik